Date: Fri, 23 Feb 2001 05:49:03 +0000 (GMT) From: Terry Lambert <tlambert@primenet.com> To: peter@netplex.com.au (Peter Wemm) Cc: tlambert@primenet.com (Terry Lambert), patl@phoenix.volant.org (Pat Lashley), freebsd-fs@FreeBSD.ORG Subject: Re: Status of nullfs Message-ID: <200102230549.WAA09447@usr05.primenet.com> In-Reply-To: <200102230242.f1N2gUf43386@mobile.wemm.org> from "Peter Wemm" at Feb 22, 2001 06:42:30 PM
next in thread | previous in thread | raw e-mail | index | archive | help
> Terry, please check your facts. This has been recently fixed. See the > VOP_GETVOBJECT() stuff. The trick though is to get it MFC'ed into > 4.x... Cool; I was unaware that this had been done. This was even the approach I advocated from a casual look at it on the CVS web. Last I had heard of this, people were still advocating alias objects. Things like this should really be talked about on the -FS list, don't you think? -- Realize that I don't run 5.x these days, since it's not distributed as a release yet. He also can't run it, if it's going to be a production environment, since it's too unstable (library stuff, etc.). > > There are a number of patches to manually enforce coherency; as > > far as I know, they don't deal with the pager path, and, since > > the coherency is explicit, the /usr vs. /null_usr problem is > > not resolved. Check the list archives to obtain the patches. > > nullfs is fully functional in -current. Backporting the changes should > be easy. It would be cool if someone could do this; I'm currently busy dealing with other fires. I stole an hour and a half tonight (more on that in a minute). From what he needs it for, you could also diddle the mount tab to implement the equivalent of a Solaris "loopback" mount, where you externalize the same FS in multiple locations. You would need to take special care with the ".." transition over the mount point, though, to make sure you didn't end up out of your jail, and in someone else's (might be done automatically; I have no idea whether the .. is handled by backing over the matching mount point (which would be "the first matching" in this case), or whether it uses the covered vnode based on the mount instance. This would be a very easy, low cost approach to making an R/O FS appear in more than one place. This would effectively be sililar to the -union option (as opposed to using a unionfs). Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200102230549.WAA09447>